ipex-llm 2.2.0b20250211__py3-none-manylinux2010_x86_64.whl → 2.2.0b20250212__py3-none-manylinux2010_x86_64.whl

ipex_llm/vllm/xpu/entrypoints/openai/cli_args.py

@@ -7,11 +7,14 @@ purposes.
 import argparse
 import json
 import ssl
-from typing import List, Optional, Sequence, Union
+from typing import List, Optional, Sequence, Union, get_args
 
 from vllm.engine.arg_utils import AsyncEngineArgs, nullable_str
+from vllm.entrypoints.chat_utils import (ChatTemplateContentFormatOption,
+                                         validate_chat_template)
 from vllm.entrypoints.openai.serving_engine import (LoRAModulePath,
                                                     PromptAdapterPath)
+from vllm.entrypoints.openai.tool_parsers import ToolParserManager
 from vllm.utils import FlexibleArgumentParser
 
 
@@ -130,10 +133,23 @@ def make_arg_parser(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
                         help="The file path to the chat template, "
                         "or the template in single-line form "
                         "for the specified model")
+    parser.add_argument(
+        '--chat-template-content-format',
+        type=str,
+        default="auto",
+        choices=get_args(ChatTemplateContentFormatOption),
+        help='The format to render message content within a chat template.'
+        '\n\n'
+        '* "string" will render the content as a string. '
+        'Example: "Hello World"\n'
+        '* "openai" will render the content as a list of dictionaries, '
+        'similar to OpenAI schema. '
+        'Example: [{"type": "text", "text": "Hello world!"}]')
     parser.add_argument("--response-role",
                         type=nullable_str,
                         default="assistant",
-                        help="The role name to return if `request.add_generation_prompt=true`.")
+                        help="The role name to return if "
+                        "`request.add_generation_prompt=true`.")
     parser.add_argument("--ssl-keyfile",
                         type=nullable_str,
                         default=None,
@@ -180,28 +196,36 @@ def make_arg_parser(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
         action="store_true",
         help="If specified, will run the OpenAI frontend server in the same "
         "process as the model serving engine.")
-
+    parser.add_argument(
+        "--enable-request-id-headers",
+        action="store_true",
+        help="If specified, API server will add X-Request-Id header to "
+        "responses. Caution: this hurts performance at high QPS.")
     parser.add_argument(
         "--enable-auto-tool-choice",
         action="store_true",
         default=False,
         help="Enable auto tool choice for supported models. Use --tool-call-parser"
-        "to specify which parser to use")
+        " to specify which parser to use")
 
+    valid_tool_parsers = ToolParserManager.tool_parsers.keys()
     parser.add_argument(
         "--tool-call-parser",
         type=str,
-        choices=["mistral", "hermes"],
+        metavar="{" + ",".join(valid_tool_parsers) + "} or name registered in "
+        "--tool-parser-plugin",
         default=None,
         help="Select the tool call parser depending on the model that you're using."
         " This is used to parse the model-generated tool call into OpenAI API "
         "format. Required for --enable-auto-tool-choice.")
 
     parser.add_argument(
-        "--load-in-low-bit",
+        "--tool-parser-plugin",
         type=str,
-        default="sym_int4",
-        help="Low-bit quantization for IPEX-LLM models")
+        default="",
+        help="Special the tool parser plugin write to parse the model-generated tool"
+        " into OpenAI API format, the name register in this plugin can be used "
+        "in --tool-call-parser.")
 
     parser = AsyncEngineArgs.add_cli_args(parser)
 
@@ -218,10 +242,35 @@ def make_arg_parser(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
         default=False,
         help="Disable FastAPI's OpenAPI schema, Swagger UI, and ReDoc endpoint"
     )
+    parser.add_argument(
+        "--enable-prompt-tokens-details",
+        action='store_true',
+        default=False,
+        help="If set to True, enable prompt_tokens_details in usage.")
+
+    parser.add_argument(
+        "--load-in-low-bit",
+        type=str,
+        default="sym_int4",
+        help="Low-bit quantization for IPEX-LLM models")
 
     return parser
 
 
+def validate_parsed_serve_args(args: argparse.Namespace):
+    """Quick checks for model serve args that raise prior to loading."""  # noqa
+    if hasattr(args, "subparser") and args.subparser != "serve":
+        return
+
+    # Ensure that the chat template is valid; raises if it likely isn't
+    validate_chat_template(args.chat_template)
+
+    # Enable auto tool needs a tool call parser to be valid
+    if args.enable_auto_tool_choice and not args.tool_call_parser:
+        raise TypeError("Error: --enable-auto-tool-choice requires "  # noqa
+                        "--tool-call-parser")
+
+
 def create_parser_for_docs() -> FlexibleArgumentParser:
     parser_for_docs = FlexibleArgumentParser(
         prog="-m vllm.entrypoints.openai.api_server")

ipex_llm/vllm/xpu/ipex_llm_v1_wrapper.py

@@ -0,0 +1,23 @@
+from vllm.logger import init_logger
+from vllm.v1.executor.ray_utils import RayWorkerWrapper
+
+
+logger = init_logger(__name__)
+
+
+class IPEXLLMV1Wrapper(RayWorkerWrapper):
+    def __init__(self, load_in_low_bit="sym_int4", *args, **kwargs) -> None:
+        super().__init__(*args, **kwargs)
+        from ipex_llm.vllm.xpu.model_convert import _ipex_llm_convert
+        _ipex_llm_convert(load_in_low_bit=load_in_low_bit)
+        self.compiled_dag_cuda_device_set = False
+
+
+def get_ipex_llm_v1_wrapper(load_in_low_bit):
+    # The reason why we not using functools.partial is that
+    # ray seems not work well with it.
+    class WrapperWithLoadBit(IPEXLLMV1Wrapper):
+        def __init__(self, *args, **kwargs) -> None:
+            super().__init__(load_in_low_bit=load_in_low_bit, *args, **kwargs)
+
+    return WrapperWithLoadBit

ipex_llm/vllm/xpu/model_convert.py

@@ -65,9 +65,14 @@ def _model_sample_convert():
 def _ipex_llm_convert(load_in_low_bit):
     from vllm.worker.xpu_model_runner import XPUModelRunner
     from ipex_llm.vllm.xpu.ipex_llm_wrapper import get_ipex_llm_wrapper
-    import vllm.executor.ray_utils as ray_utils
+    from ipex_llm.vllm.xpu.ipex_llm_v1_wrapper import get_ipex_llm_v1_wrapper
+    import vllm.executor.ray_utils as ray_utils_v0
+    import vllm.v1.executor.ray_utils as ray_utils_v1
+    from vllm.v1.worker.gpu_model_runner import GPUModelRunner
     setattr(XPUModelRunner, "load_model", get_load_function(load_in_low_bit))
-    setattr(ray_utils, "RayWorkerWrapper", get_ipex_llm_wrapper(load_in_low_bit))
+    setattr(GPUModelRunner, "load_model", get_load_function(load_in_low_bit))
+    setattr(ray_utils_v0, "RayWorkerWrapper", get_ipex_llm_wrapper(load_in_low_bit))
+    setattr(ray_utils_v1, "RayWorkerWrapper", get_ipex_llm_v1_wrapper(load_in_low_bit))
 
 
 def get_load_function(low_bit):
@@ -77,19 +82,16 @@ def get_load_function(low_bit):
         # from vllm.utils import measure_device_memory
         from vllm.utils import DeviceMemoryProfiler
         with DeviceMemoryProfiler() as m:
+            from dataclasses import replace
+            new_device_config = DeviceConfig("cpu")
+            new_vllm_config = replace(self.vllm_config, device_config=new_device_config)
             self.model = get_model(
-                model_config=self.model_config,
-                device_config=DeviceConfig("cpu"),
-                load_config=self.load_config,
-                lora_config=self.lora_config,
-                parallel_config=self.parallel_config,
-                scheduler_config=self.scheduler_config,
-                cache_config=self.cache_config,
+                vllm_config=new_vllm_config
             )
-            if "qwen" in self.model_config.model.lower() or \
-                    "baichuan" in self.model_config.model.lower() or \
-                    "codegeex4-all" in self.model_config.model.lower() or \
-                    "chatglm" in self.model_config.model.lower():
+            if "qwen" in self.vllm_config.model_config.model.lower() or \
+                    "baichuan" in self.vllm_config.model_config.model.lower() or \
+                    "codegeex4-all" in self.vllm_config.model_config.model.lower() or \
+                    "chatglm" in self.vllm_config.model_config.model.lower():
                 self.model.apply(padding_mlp)
             from ipex_llm import optimize_model
             import os
@@ -99,18 +101,22 @@ def get_load_function(low_bit):
                 modules = ["35.mlp", "36.mlp", "37.mlp", "38.mlp", "39.mlp"]
             else:
                 modules = None
-            if "minicpm" in self.model_config.model.lower():
+            if "minicpm" in self.vllm_config.model_config.model.lower():
                 modules = ["vpm", "resampler"]
             # only for minicpm_2_6
-            if "minicpm-v" in self.model_config.model.lower():
+            if "minicpm-v" in self.vllm_config.model_config.model.lower():
                 from ipex_llm.transformers.models.minicpmv import merge_qkv
                 self.model.vpm.apply(merge_qkv)
-            if "internvl2" in self.model_config.model.lower():
+            if "internvl2" in self.vllm_config.model_config.model.lower():
                 modules = ["vision_model", "mlp1"]
-            optimize_model(self.model, low_bit=low_bit, torch_dtype=self.model_config.dtype,
+            if "deepseek-v2" in self.vllm_config.model_config.model.lower():
+                modules = ["down_proj"]
+            optimize_model(self.model,
+                           low_bit=low_bit,
+                           torch_dtype=self.vllm_config.model_config.dtype,
                            modules_to_not_convert=modules)
-            self.model = self.model.to(device=self.device_config.device,
-                                       dtype=self.model_config.dtype)
+            self.model = self.model.to(device=self.vllm_config.device_config.device,
+                                       dtype=self.vllm_config.model_config.dtype)
 
         self.model_memory_usage = m.consumed_memory
         logger = init_logger(__name__)

{ipex_llm-2.2.0b20250211.dist-info → ipex_llm-2.2.0b20250212.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ipex-llm
-Version: 2.2.0b20250211
+Version: 2.2.0b20250212
 Summary: Large Language Model Develop Toolkit
 Home-page: https://github.com/intel-analytics/ipex-llm
 Author: BigDL Authors
@@ -27,7 +27,7 @@ Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine
 Requires-Dist: torch ==2.1.2+cpu ; (platform_system == "Linux") and extra == 'all'
 Requires-Dist: torch ==2.1.2 ; (platform_system == "Windows") and extra == 'all'
 Provides-Extra: cpp
-Requires-Dist: bigdl-core-cpp ==2.6.0b20250211 ; extra == 'cpp'
+Requires-Dist: bigdl-core-cpp ==2.6.0b20250212 ; extra == 'cpp'
 Requires-Dist: setuptools ; extra == 'cpp'
 Requires-Dist: onednn-devel ==2025.0.1 ; (platform_system == "Windows") and extra == 'cpp'
 Requires-Dist: onednn ==2025.0.1 ; (platform_system == "Windows") and extra == 'cpp'
@@ -60,7 +60,7 @@ Requires-Dist: transformers ==4.40.0 ; extra == 'npu'
 Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'npu'
 Requires-Dist: torch ==2.1.2+cpu ; (platform_system == "Linux") and extra == 'npu'
 Requires-Dist: torch ==2.1.2 ; (platform_system == "Windows") and extra == 'npu'
-Requires-Dist: bigdl-core-npu ==2.6.0b20250211 ; (platform_system == "Windows") and extra == 'npu'
+Requires-Dist: bigdl-core-npu ==2.6.0b20250212 ; (platform_system == "Windows") and extra == 'npu'
 Provides-Extra: serving
 Requires-Dist: py-cpuinfo ; extra == 'serving'
 Requires-Dist: fschat[model_worker,webui] ==0.2.36 ; extra == 'serving'
@@ -80,9 +80,9 @@ Requires-Dist: setuptools <70.0.0 ; extra == 'xpu'
 Requires-Dist: torch ==2.1.0a0 ; extra == 'xpu'
 Requires-Dist: torchvision ==0.16.0a0 ; extra == 'xpu'
 Requires-Dist: intel-extension-for-pytorch ==2.1.10+xpu ; extra == 'xpu'
-Requires-Dist: bigdl-core-xe-21 ==2.6.0b20250211 ; extra == 'xpu'
-Requires-Dist: bigdl-core-xe-batch-21 ==2.6.0b20250211 ; extra == 'xpu'
-Requires-Dist: bigdl-core-xe-addons-21 ==2.6.0b20250211 ; extra == 'xpu'
+Requires-Dist: bigdl-core-xe-21 ==2.6.0b20250212 ; extra == 'xpu'
+Requires-Dist: bigdl-core-xe-batch-21 ==2.6.0b20250212 ; extra == 'xpu'
+Requires-Dist: bigdl-core-xe-addons-21 ==2.6.0b20250212 ; extra == 'xpu'
 Provides-Extra: xpu-2-1
 Requires-Dist: py-cpuinfo ; extra == 'xpu-2-1'
 Requires-Dist: protobuf ; extra == 'xpu-2-1'
@@ -97,9 +97,9 @@ Requires-Dist: setuptools <70.0.0 ; extra == 'xpu-2-1'
 Requires-Dist: torch ==2.1.0a0 ; extra == 'xpu-2-1'
 Requires-Dist: torchvision ==0.16.0a0 ; extra == 'xpu-2-1'
 Requires-Dist: intel-extension-for-pytorch ==2.1.10+xpu ; extra == 'xpu-2-1'
-Requires-Dist: bigdl-core-xe-21 ==2.6.0b20250211 ; extra == 'xpu-2-1'
-Requires-Dist: bigdl-core-xe-batch-21 ==2.6.0b20250211 ; extra == 'xpu-2-1'
-Requires-Dist: bigdl-core-xe-addons-21 ==2.6.0b20250211 ; extra == 'xpu-2-1'
+Requires-Dist: bigdl-core-xe-21 ==2.6.0b20250212 ; extra == 'xpu-2-1'
+Requires-Dist: bigdl-core-xe-batch-21 ==2.6.0b20250212 ; extra == 'xpu-2-1'
+Requires-Dist: bigdl-core-xe-addons-21 ==2.6.0b20250212 ; extra == 'xpu-2-1'
 Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-2-1'
 Requires-Dist: dpcpp-cpp-rt ==2024.0.2 ; (platform_system == "Windows") and extra == 'xpu-2-1'
 Requires-Dist: mkl-dpcpp ==2024.0.0 ; (platform_system == "Windows") and extra == 'xpu-2-1'
@@ -117,7 +117,7 @@ Requires-Dist: setuptools ; extra == 'xpu-2-6'
 Requires-Dist: torch ==2.6.0+xpu ; extra == 'xpu-2-6'
 Requires-Dist: torchvision ==0.21.0+xpu ; extra == 'xpu-2-6'
 Requires-Dist: torchaudio ==2.6.0+xpu ; extra == 'xpu-2-6'
-Requires-Dist: bigdl-core-xe-all ==2.6.0b20250211 ; extra == 'xpu-2-6'
+Requires-Dist: bigdl-core-xe-all ==2.6.0b20250212 ; extra == 'xpu-2-6'
 Requires-Dist: onednn-devel ==2025.0.1 ; extra == 'xpu-2-6'
 Requires-Dist: onednn ==2025.0.1 ; extra == 'xpu-2-6'
 Requires-Dist: dpcpp-cpp-rt ==2025.0.2 ; extra == 'xpu-2-6'
@@ -133,9 +133,9 @@ Requires-Dist: tokenizers ==0.15.2 ; extra == 'xpu-arc'
 Requires-Dist: accelerate ==0.23.0 ; extra == 'xpu-arc'
 Requires-Dist: tabulate ; extra == 'xpu-arc'
 Requires-Dist: setuptools ; extra == 'xpu-arc'
-Requires-Dist: bigdl-core-xe-23 ==2.6.0b20250211 ; extra == 'xpu-arc'
-Requires-Dist: bigdl-core-xe-batch-23 ==2.6.0b20250211 ; extra == 'xpu-arc'
-Requires-Dist: bigdl-core-xe-addons-23 ==2.6.0b20250211 ; extra == 'xpu-arc'
+Requires-Dist: bigdl-core-xe-23 ==2.6.0b20250212 ; extra == 'xpu-arc'
+Requires-Dist: bigdl-core-xe-batch-23 ==2.6.0b20250212 ; extra == 'xpu-arc'
+Requires-Dist: bigdl-core-xe-addons-23 ==2.6.0b20250212 ; extra == 'xpu-arc'
 Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-arc'
 Requires-Dist: torch ==2.3.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-arc'
 Requires-Dist: torchvision ==0.18.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-arc'
@@ -156,9 +156,9 @@ Requires-Dist: tokenizers ==0.15.2 ; extra == 'xpu-arl'
 Requires-Dist: accelerate ==0.23.0 ; extra == 'xpu-arl'
 Requires-Dist: tabulate ; extra == 'xpu-arl'
 Requires-Dist: setuptools ; extra == 'xpu-arl'
-Requires-Dist: bigdl-core-xe-23 ==2.6.0b20250211 ; extra == 'xpu-arl'
-Requires-Dist: bigdl-core-xe-batch-23 ==2.6.0b20250211 ; extra == 'xpu-arl'
-Requires-Dist: bigdl-core-xe-addons-23 ==2.6.0b20250211 ; extra == 'xpu-arl'
+Requires-Dist: bigdl-core-xe-23 ==2.6.0b20250212 ; extra == 'xpu-arl'
+Requires-Dist: bigdl-core-xe-batch-23 ==2.6.0b20250212 ; extra == 'xpu-arl'
+Requires-Dist: bigdl-core-xe-addons-23 ==2.6.0b20250212 ; extra == 'xpu-arl'
 Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-arl'
 Requires-Dist: torch ==2.3.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-arl'
 Requires-Dist: torchvision ==0.18.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-arl'
@@ -179,9 +179,9 @@ Requires-Dist: tokenizers ==0.15.2 ; extra == 'xpu-lnl'
 Requires-Dist: accelerate ==0.23.0 ; extra == 'xpu-lnl'
 Requires-Dist: tabulate ; extra == 'xpu-lnl'
 Requires-Dist: setuptools ; extra == 'xpu-lnl'
-Requires-Dist: bigdl-core-xe-23 ==2.6.0b20250211 ; extra == 'xpu-lnl'
-Requires-Dist: bigdl-core-xe-batch-23 ==2.6.0b20250211 ; extra == 'xpu-lnl'
-Requires-Dist: bigdl-core-xe-addons-23 ==2.6.0b20250211 ; extra == 'xpu-lnl'
+Requires-Dist: bigdl-core-xe-23 ==2.6.0b20250212 ; extra == 'xpu-lnl'
+Requires-Dist: bigdl-core-xe-batch-23 ==2.6.0b20250212 ; extra == 'xpu-lnl'
+Requires-Dist: bigdl-core-xe-addons-23 ==2.6.0b20250212 ; extra == 'xpu-lnl'
 Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-lnl'
 Requires-Dist: torch ==2.3.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-lnl'
 Requires-Dist: torchvision ==0.18.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-lnl'

{ipex_llm-2.2.0b20250211.dist-info → ipex_llm-2.2.0b20250212.dist-info}/RECORD

@@ -94,7 +94,7 @@ ipex_llm/serving/fastchat/tgi_api_protocol.py,sha256=brT3k3-V0NJrU4fRqUwWjC0O3iO
 ipex_llm/serving/fastchat/tgi_api_server.py,sha256=agNTAEiZPSuj3dEdIdYKwkoY0cXOUDX06DiM9VP2knQ,24418
 ipex_llm/serving/fastchat/vllm_worker.py,sha256=ZLz2Q9GxJO6r_LOiP6epgCRjBGk-K4EB1SNEWSJp5DA,11091
 ipex_llm/transformers/__init__.py,sha256=pJHs6GZXHIObVE4BUCuej-6BKBZZg9pYWKPrkhWSfB4,1192
-ipex_llm/transformers/convert.py,sha256=42qHApc3hoL38ldQXTv9BkDVQ0Zb-xGJYA2DWYFzwMg,100795
+ipex_llm/transformers/convert.py,sha256=t-2tMK9ktT0GGszoN65gQfVtBvUV1Yka7uyp4CUFGok,100851
 ipex_llm/transformers/convert_ipex.py,sha256=_nSnUTQy-yfkKaqGdqnBdWztZf3NGmnbZ0TKaDrF4X4,14617
 ipex_llm/transformers/embedding.py,sha256=bdgk59DvD4ZZyxRzewXOR7g56nThgO6uhIwk8QL7f-s,9299
 ipex_llm/transformers/kv.py,sha256=k4TU18LlA-Sbq9WNNQnfuzu3RSFBwFhmaV3BcGN5bAo,19191
@@ -162,6 +162,7 @@ ipex_llm/transformers/models/gptbigcode.py,sha256=cP1_qGWoa43R2WacAMblShjku4Qupc
 ipex_llm/transformers/models/gptneox.py,sha256=loRh1x_5S6BCeOr_s5xr-N_1SQHL3Y5IiUBAEyoMUqQ,6172
 ipex_llm/transformers/models/internlm.py,sha256=JZFrI2HXsIAfM-6pA2RO0wcXopOliC1FggLMzNzaDZ4,17404
 ipex_llm/transformers/models/internvl.py,sha256=Vx0vENIEQLX2M6P398mw5TOhpks0U8xf8rtRQvy94go,8154
+ipex_llm/transformers/models/janus.py,sha256=0URo2NC8_2CGaOl3CiVB3IFTVsYyplMFgjBJdPDNBsY,1509
 ipex_llm/transformers/models/llama.py,sha256=rqrNjuZb_jeb9MKx0z-FSVoGx8YDBxQzPJ9ZUvYhgx0,9138
 ipex_llm/transformers/models/minicpm.py,sha256=eaPNVNrep0_xGoELhZd886ff0ceoKqB6cusdAhd52eE,10145
 ipex_llm/transformers/models/minicpm3.py,sha256=11cYl8KM2hoIJNMAOZMxiwCu6dMhup9ric_OEn8-VrQ,9363
@@ -182,7 +183,7 @@ ipex_llm/transformers/models/rwkv5.py,sha256=OkRNj1pCAZg1z2Fw-I0DEnxLEdZyPeRSQ6m
 ipex_llm/transformers/models/sd.py,sha256=VvHV5u-0k2MgHu3NL9113hPj7DgfxqctuKzEEeNfRDU,5981
 ipex_llm/transformers/models/stablelm.py,sha256=fj-XtOnR6kggnFUQTMPCOOzolkPztN06WAv8QW-XRnI,7054
 ipex_llm/transformers/models/starcoder2.py,sha256=ONKvD7JCkRM0DI-R56x28QFBJ7CjD5hOZBQ_3WfOcNk,6626
-ipex_llm/transformers/models/utils.py,sha256=qI5ln8SQGTvR_IyxFkoZhefgOErnXUnJrifIyhiqT9c,14753
+ipex_llm/transformers/models/utils.py,sha256=c3hh0YDHE-Qg7SQBXhnNXf85Nx7jopZFfa1KS-Pe6kQ,14734
 ipex_llm/transformers/models/whisper.py,sha256=ju3WP8Eq-KvD7kb3Qy51r4FOfSX3NBxfp5RBcq__gzc,4241
 ipex_llm/transformers/models/yuan.py,sha256=JYAn_ZaSGK0NBJLEIxCACfAq084a66GFJkdd5NbpmMA,7732
 ipex_llm/transformers/npu_models/__init__.py,sha256=ulEUGLjaP48LCrVeury3UxLjXxKzRi0UpSG4bYu-7f8,585
@@ -250,17 +251,18 @@ ipex_llm/vllm/cpu/entrypoints/api_server.py,sha256=RDe3GrGMw47c7dVZL2IMmTHtGegnz
 ipex_llm/vllm/cpu/entrypoints/openai/api_server.py,sha256=l0tHnCaD96WD2k3zLTrYtLrcigf1eqYvszs79fYgoK8,29812
 ipex_llm/vllm/cpu/entrypoints/openai/cli_args.py,sha256=hB398yYtKauASRzevctScdbFIjiiSGMAe1bwEuIHrhY,10893
 ipex_llm/vllm/xpu/__init__.py,sha256=zBSG6nzrVF5QnpR6_f7kPhBFeowTE9gaZ7D5m98E7_w,585
+ipex_llm/vllm/xpu/ipex_llm_v1_wrapper.py,sha256=pd939vFomKIg9Qn2NO4u0OF6hPgvQpqcfJSxqBzcqhA,825
 ipex_llm/vllm/xpu/ipex_llm_wrapper.py,sha256=_CbhvBuf_KPnmLfngYKtJl5gPAHVsG2mWth3wSeaH3M,892
-ipex_llm/vllm/xpu/model_convert.py,sha256=tDgaOliyBJdpcCctFNlMd9RK8fA7usLQi2RadgOj--A,7599
+ipex_llm/vllm/xpu/model_convert.py,sha256=0lfzHhCtTVSZZdlqXAOAEtK3FsouqEheSGr9MjpWz38,8137
 ipex_llm/vllm/xpu/engine/__init__.py,sha256=pY_CpyuZd72fr6s32ejeKHKFW0K4vUU2rzZjWqm5YiA,783
-ipex_llm/vllm/xpu/engine/engine.py,sha256=k4-D27WS_Gk3mA--w3HWAjPjb4Aiu043MVPi0ZoAUBc,5984
-ipex_llm/vllm/xpu/entrypoints/openai/api_server.py,sha256=GshTZFB8e4PWvqckfbmTOU6b0oLkNn7A-vzLuG9--j8,21544
-ipex_llm/vllm/xpu/entrypoints/openai/cli_args.py,sha256=2rENA2ucynMaIjiZBEh2ez1o5vR32GaP514t39CD7KM,8676
-ipex_llm-2.2.0b20250211.data/scripts/ipex-llm-init,sha256=fLQsT2dRL6H5bThb4GuIWotAuqoLsIxFwA-0c2qmaO8,6672
-ipex_llm-2.2.0b20250211.data/scripts/llm-chat,sha256=TdUnUmNapzuoe1c8IzrdVOQwWEg8IqsMSBRlOD3daZM,2249
-ipex_llm-2.2.0b20250211.data/scripts/llm-cli,sha256=RXGPlLElHxcKzoUxljEMBIAXbzCDysXL-Nxw-xF-7LU,2457
-ipex_llm-2.2.0b20250211.dist-info/METADATA,sha256=bXwvxJIBVIayYIeP1lVUyFJXiMqzVppGcvwttejTvOk,12369
-ipex_llm-2.2.0b20250211.dist-info/WHEEL,sha256=PPJcBMAZibF_2GFE9NmOJGqiaSMPiNFbJd6QaJjdA6Y,109
-ipex_llm-2.2.0b20250211.dist-info/entry_points.txt,sha256=TiUyBB2MRmfF3ko-pyAEzqeBCRnyhu27bNOAsWPp3e8,61
-ipex_llm-2.2.0b20250211.dist-info/top_level.txt,sha256=CGCMHM-SyqUabU4h8RqJ2KTYckQUO3LvIWwmUQ6Qbzw,9
-ipex_llm-2.2.0b20250211.dist-info/RECORD,,
+ipex_llm/vllm/xpu/engine/engine.py,sha256=ClvTHz_w52IU_vUw-LhotLduCMu0zPl9kebnQEKQq4A,9494
+ipex_llm/vllm/xpu/entrypoints/openai/api_server.py,sha256=OScQFmAdrd-PyEyQ2Mr8fJrk_lC6MHqPraphVk-NDzw,33140
+ipex_llm/vllm/xpu/entrypoints/openai/cli_args.py,sha256=hB398yYtKauASRzevctScdbFIjiiSGMAe1bwEuIHrhY,10893
+ipex_llm-2.2.0b20250212.data/scripts/ipex-llm-init,sha256=fLQsT2dRL6H5bThb4GuIWotAuqoLsIxFwA-0c2qmaO8,6672
+ipex_llm-2.2.0b20250212.data/scripts/llm-chat,sha256=TdUnUmNapzuoe1c8IzrdVOQwWEg8IqsMSBRlOD3daZM,2249
+ipex_llm-2.2.0b20250212.data/scripts/llm-cli,sha256=RXGPlLElHxcKzoUxljEMBIAXbzCDysXL-Nxw-xF-7LU,2457
+ipex_llm-2.2.0b20250212.dist-info/METADATA,sha256=VFWO-HzZd_Ecx1J6HceoMwKM-u8x4aYVet1wDKSO7Ig,12369
+ipex_llm-2.2.0b20250212.dist-info/WHEEL,sha256=PPJcBMAZibF_2GFE9NmOJGqiaSMPiNFbJd6QaJjdA6Y,109
+ipex_llm-2.2.0b20250212.dist-info/entry_points.txt,sha256=TiUyBB2MRmfF3ko-pyAEzqeBCRnyhu27bNOAsWPp3e8,61
+ipex_llm-2.2.0b20250212.dist-info/top_level.txt,sha256=CGCMHM-SyqUabU4h8RqJ2KTYckQUO3LvIWwmUQ6Qbzw,9
+ipex_llm-2.2.0b20250212.dist-info/RECORD,,